Announcement

Collapse
No announcement yet.

Mmemtest Pro crashes PC

Collapse
X
Collapse
 
  • Page of 1
  • Filter
  • Time
  • Show
Clear All
new posts
Previous template Next
  • #1

    Mmemtest Pro crashes PC

    I hunting a source of OS corruptions as mentioned in a side comment here. On the first test, after leaving PC unattended for 10+ h, when returned to it I saw black sceens, Asus x570 Hero Wifi MB showing code 00 (Not Used) and power, reset buttons not reacing. Not sure what happened with the first log, may have deleted it accidentally, but I managed to get get this behaviour again and did get the log.

    It happened in Test 6 (Block move) some 2 hours into the test.

    Hardware is
    MB - Asus x570 Hero Wifi (fw ver 4402)
    CPU - Ryzen 9 5950X
    RAM - 4 x Corsair CMK32GX4M2D3200C16 (16 GB) running at 3200 MHz

    End of the log contained this:
    2023-04-27 22:06:30 - Get_AMD_19_CurTmp: AMD Ryzen 9 5950X 16-Core Processor - NB_SMN_INDEX_0=0x00059800 NB_SMN_DATA_0=0x6DFB0000
    2023-04-27 22:06:30 - Get_AMD_19_CurTmp: Setting NB_SMN_INDEX_0 to 0x00059800
    2023-04-27 22:06:31 - Get_AMD_19_CurTmp: NB_SMN_DATA_0=0x6DFB0000
    2023-04-27 22:06:31 - Get_AMD_19_CurTmp: 0x6DFB0000, 879 (RANGE_SEL=1)
    2023-04-27 22:06:31 - Get_AMD_19_CurTmp: usingBlanketOffset 60
    2023-04-27 22:06:31 - Get_AMD_19_CurTmp: iTrueTmp 60
    2023-04-27 22:06:31 - Current CPU temperature: 60C
    2023-04-27 22:06:38 - MtSupportRunAllTests - Test execution time: 785.825s (Test 5 cumulative error count: 0, buffer full count: 0)
    2023-04-27 22:06:38 - Running test #6 (Test 6 [Block move, 64-byte blocks])
    2023-04-27 22:06:38 - MtSupportRunAllTests - Setting random seed to 0xB0AD82AA
    2023-04-27 22:06:38 - MtSupportRunAllTests - Start time: 10126344 ms
    2023-04-27 22:06:38 - MtSupportRunAllTests - Enabling memory cache for test
    2023-04-27 22:06:38 - MtSupportRunAllTests - Enabling memory cache complete
    2023-04-27 22:06:38 - Start memory range test (0x0 - 0x1030000000)

    I tried selecting just Test 6 and running that again, then it crashed immediately. Tried again, crashed after about 30 minutes.
    All 3 logs end with "Start memory range test"

    Clearly something is wrong, but I am no closer to finding out is it RAM and which stick or memory controller in CPU.

    Why would memtest crash the machine this way? Any tips what is more likely to be the problem?

    Tomorrow I am getting my hands on Crucial MTA18ASF4G72AZ-3G2R 32GB ECC RAM to try identify wether CPU or MB is the problem.
    Also getting Ryzen 7 PRO 4750G CPU to see if that changes anything.
    ​
    Tags: None
  • #2
    Started tesing with individual RAM sticks. Managed to get the issue at least with two sticks. Seems unlikely that two sticks would fail. I guess that leaves CPU or MB slots or fimware as the main suspects.

    Comment

    • #3
      Why would memtest crash the machine this way?
      At the risk of stating the obvious: It isn't Memtest86 crashing the machine. It is the machine crashing Memtest86.
      If the hardware doesn't work to some degree, then the software can't work either.

      Seems unlikely that two sticks would fail
      True. But from time to time there is a design failure, or a config failure. e.g. BIOS is setting timings / voltage in such as way that none of the sticks will work.
      No easy solution for an end user except trial and error swapping hardware.

      See also this page
      https://www.memtest86.com/troubleshooting.htm


      Comment

      • #4
        Yes, of course machine was crashing memtest. I meant what it could run into to crash so badly?


        All sticks were failing individually, even tried different MB slots. Reflashed UEFI firmware and then it worked. When comparing with previous settings, noticed they had PBO fMAX Enhancer enabled which I don't usually do. must have had accidentally enabled when turning on PBO after system SSD upgrade recently. With it, Test 6 seems fail within 3 passes. Without that have ran up to 15 until stopped the test.

        Now have to test those OS backup and recovery setups that were failing very easily. Hopefully that was the root of corruptions.

        Comment

        • #5
          PBO fMAX Enhancer
          Whatever this setting is or does, it has a very dumb unintuitive name. And it crashes your system as well.
          Their are really too many very poorly documented and poorly tested settings in modern BIOS. Many of which interact with each other to create a confusing overlapping mess. Only 0.1% of users want to take the time to tweak all this (and deal with the instability). The other 99.9% just want a stable fast computer out of the box.
          This is a desperate (failed) search by ASUS for exclusive BIOS features to try and differentiate their product in the market place.

          Comment

          Previous template Next
          Working...
          X